membranes

Article
The Selection of Efficient Antiscalant for RO Facility, Control
of Its Quality and Evaluation of the Economical Efficiency of
Its Application
Dmitry Spitsov, Htet Zaw Aung and Alexei Pervov *

Department of Water Supply and Wastewater Treatment, Moscow State University of Civil Engineering, 26,
Yaroslaskoye Highway, 129337 Moscow, Russia
* Correspondence: [email protected]

Abstract: Adsorption of polymeric inhibitor molecules to calcium carbonate crystal surface was inves-
tigated. Inhibiting efficiencies of phosphonic acid-based antiscalants are dependent on the amount of
adsorbed material on the growing crystal surface. A strong antiscalant even at a small dose provides
the necessary rate of adsorption. Comparison of two phosphonic-based antiscalants was made both in
laboratory and industrial conditions. A distinguishing feature of the strong antiscalant is the presence of
aminotris (metylene-diphosphonic acid) ATMP. Experimental dependencies of antiscalant adsorption
rates on the antiscalant dosage values were determined. Emphasis is given to the use of nanofiltration
membranes that possess lower scaling propensities. Modernization is presented to reduce operational
costs due to antiscalant and nanofiltration membranes. The main conclusion is that control of scaling
should be implemented together with the use of nanofiltration membranes.

Keywords: reverse osmosis; calcium carbonate; scaling mechanism; phosphonic-based antiscalants;

inhibitors; crystallization mechanism; membrane cleanings; concentrate disposal

Citation: Spitsov, D.; Aung, H.Z.;

Pervov, A. The Selection of Efficient
1. Introduction
Antiscalant for RO Facility, Control of
Its Quality and Evaluation of the Application of antiscalants to control sparingly soluble salts that deposit in reverse
Economical Efficiency of Its osmosis membrane units has become one of the most important and significant issues in
Application. Membranes 2023, 13, 85. reverse osmosis practice [1–4]. As was already discussed [5–7], proposed antiscalants with
https://doi.org/10.3390/ new unknown brands require evaluation of their efficiencies. but antiscalant testing results
membranes13010085 provided by different researchers often do not provide real conditions for crystal nucleation
that exist in membrane industrial devices. This often does not enable us to efficiently apply
Academic Editors: Semyon
new antiscalants and evaluate their behavior [8–10]. Often laboratory testing provides
Mareev, Dmitrii Butylskii and
Mikhail Porozhnyy
inconsistent results that cannot be transferred into practice [1,8]. In previously published
articles several types of phosphonic and acrylic inhibitors have been tested to compare their
Received: 28 November 2022 effectiveness. Comparison of adsorption rates and adsorption capacities was conducted.
Revised: 22 December 2022 Now we want to demonstrate the benefits of using reduced doses.
Accepted: 28 December 2022 Phosphonates or phosphonic acid-based antiscalants are still recognized as leading
Published: 9 January 2023 products in the reverse osmosis desalination market [5,10]. This leadership is still main-
tained despite the diversity of their inhibiting abilities. A lot of research was undertaken to
develop and introduce into practice “green” antiscalants and a variety of different classes of
antiscalants: acrylic acid antiscalants, polyaspartate-based antiscalants, organophosphates
Copyright: © 2023 by the authors.
and others [9–12]. This is attributed to the reliability and efficiency of phosphonic-based
Licensee MDPI, Basel, Switzerland.
This article is an open access article
antiscalants over other classes of inhibitors. Yet belonging to the class of phosphonates
distributed under the terms and
does not guarantee high efficiency. Operational experience has demonstrated that different
conditions of the Creative Commons antiscalants even among phosphonates demonstrate different efficiencies that results in cer-
Attribution (CC BY) license (https:// tain operational costs, losses and damages. This article aims to demonstrate the differences
creativecommons.org/licenses/by/ in antiscaling efficiencies of different phosphonates. As will be discussed, the difference
4.0/).

Membranes 2023, 13, 85. https://doi.org/10.3390/membranes13010085 https://www.mdpi.com/journal/membranes

Membranes 2023, 13, x FOR PEER REVIEW 2 of 17

phosphonates does not guarantee high efficiency. Operational experience has demon-
Membranes 2023, 13, 85 strated that different antiscalants even among phosphonates demonstrate different 2effi- of 15
ciencies that results in certain operational costs, losses and damages. This article aims to
demonstrate the differences in antiscaling efficiencies of different phosphonates. As will
be discussed, the difference in phosphonates’ behavior is mainly attributed to the amount
in phosphonates’ behavior is mainly attributed to the amount of aminotris (methylene-
of aminotris(methylene-phosphonic acid)—ATMP—that is formed during antiscalant
phosphonic acid)—ATMP—that is formed during antiscalant synthesis. Figure 1 demon-
synthesis. Figure 1 demonstrates spectrum results of nuclei-magnetic resonance analysis
strates spectrum results of nuclei-magnetic resonance analysis that show the ratio of ATMP
that show the ratio of ATMP and nitrilo-trimethyl-phosphonic acid (NTMP). The “Jurby-
and nitrilo-trimethyl-phosphonic acid (NTMP). The “Jurby-Soft” inhibitor consists entirely
Soft” inhibitor consists entirely of NTMP and does not contain ATMP. This antiscalant
of NTMP and does not contain ATMP. This antiscalant demonstrated the lowest efficiency.
demonstrated the lowest efficiency. Different phosphonic-based antiscalants demonstrate
Different phosphonic-based antiscalants demonstrate high efficiencies, but among them
high efficiencies, but among them “Aminat-K” demonstrates maximum value of
“Aminat-K” demonstrates maximum value of NTMP/ATMP ratio (equal to 5:1), corre-
NTMP/ATMP ratio (equal to 5:1), corresponding to its maximal efficiency [1,12]. The pre-
sponding to its maximal efficiency [1,12]. The present article aims to provide experimental
sent article aims to provide experimental proof of this superiority and to demonstrate how
proof of this superiority and to demonstrate how antiscalant efficiency ensures reduction
antiscalant efficiency
of operational costs. ensures reduction
These results are of operational
very useful forcosts. These results are
RO management whenvery
theuseful
most
for RO management when the most efficient product should be selected and purchased.
efficient product should be selected and purchased. This method is also very helpful when
This methodofisthealso
the quality very helpful
supplied reagentwhen
shouldthe
be quality
checkedofandthethe
supplied reagent
input quality should
control be
of the
checked and the input quality
product should be arranged. control of the product should be arranged.

Figure
Figure1.1.Laboratory
Laboratorytest
testunit
unit flow
flow diagram:
diagram: 1—source watertank;
1—source water tank;2—pump;
2—pump;3—membrane
3—membraneelement
element
in
in the pressure vessel; 4—filtrate tank; 5—heat exchanger; 6—manometer; 7–9—flow meters;
the pressure vessel; 4—filtrate tank; 5—heat exchanger; 6—manometer; 7–9—flow meters; 10—bypass 10—
bypass valve; 11—valve
valve; 11—valve for adjusting
for adjusting the flowthe
of flow ofwater;
source source12—valve
water; 12—valve for adjusting
for adjusting the working
the working pressure
pressure and concentrate flow; 13—valve for adjusting the flow of cooling water; 14, 15—samplers.
and concentrate flow; 13—valve for adjusting the flow of cooling water; 14, 15—samplers.

At
At the
the present
present time,
time, interest
interest in in using
using reverse
reverse osmosis
osmosis in in drinking
drinking water
water supply
supply isis
growing.
growing. Many
Many newnew suppliers
suppliersofofmembrane
membrane products
products andand service
service chemicals
chemicals are con-
are constantly
stantly appearing on the market. The tender principle of procurement
appearing on the market. The tender principle of procurement leads to the situation leads to the situa-
that
tion that a low quality and inefficient product is supplied. The problem
a low quality and inefficient product is supplied. The problem is the correct formulation is the correct for-
mulation of the antiscalant
of the antiscalant compositioncomposition requirements.
requirements. This to
This is due is an
dueincorrectly
to an incorrectly for-
formulated
mulated requirement
requirement to supply to supply
a mixture a mixture of sodium
of sodium saltsphosphonic
salts and and phosphonic acids,acids,
whichwhich is
is not
not enough
enough to solve
to solve the the problem.
problem. TheThe initial
initial choice
choice of the
of the antiscalant
antiscalant should
should be based
be based on
on the
the results
results of the
of the analysis,
analysis, butbut later,
later, whenwhenlargelarge amounts
amounts of antiscalants
of antiscalants areare purchased,
purchased, we
we should
should analyze
analyze antiscalant
antiscalant samples’
samples’ nuclear
nuclear magnetic
magnetic resonance
resonance spectrums
spectrums to ensure
to ensure that
the required
that product
the required is used.
product In this
is used. In article, the authors
this article, makemake
the authors an attempt to demonstrate
an attempt to demon-a
comparison
strate of antiscalant
a comparison efficiency
of antiscalant with an account
efficiency with an of the reduction
account of operational
of the reduction costs
of opera-
due tocosts
tional the dosing
due to of
thehigh-quality product in the
dosing of high-quality feed water.
product in the feed water.

2. Experimental Part: Materials and Equipment

The goal of the experiments was to demonstrate the adsorption characteristics of
different antiscalants and to connect adsorption ability and antiscaling efficiency of different
products with the purpose of developing guidelines to reduce operational costs. The
experimental program included:
1. Determination of calcium carbonate crystallization rates depending on antiscalant
dose and membrane type.
Membranes 2023, 13, 85 3 of 15

2. Evaluation of antiscalant adsorption rates for different doses.

3. Dependencies of adsorption rates on the antiscalant type and its dose.
The laboratory test unit flow diagram is shown in Figure 1. The evaluation of scaling
rates was conducted in accordance with the test procedure developed previously and
described in a number of publications [1,12]. The test procedure was based on operation of
the test unit in recirculation mode whereby concentrate of membrane module was returned
back to the feed water tank and permeate was constantly withdrawn from the system
and collected in a separate tank [1]. Underground water was used in the experiments.
Samples were taken from water intake wells at reverse osmosis water treatment plants in the
Moscow area. The TDS value of the ground water was 700 ppm, calcium concentration was
5.6 milliequivalents per liter; alkalinity was 120 ppm; chloride concentration was 210 ppm;
sulphate concentration was 288 ppm; pH was 7.2. Water composition of samples taken
from three different wells is presented in Table 1. The test procedure included circulation
of the feed water in the membrane unit and evaluation of scaling rates using the method of
mass balance.

Table 1. Ground water composition in different intakes (Moscow region).

Component Intake 1 Intake 2 Intake 3

1 Ca2+ 4.4 3.9 3.5
2 Mg2+ 2.3 1.4 2.5
3 SO4 2− 0.63 0.2 0.7
4 Cl− 0.29 0.3 0.15
5 HCO3 − 6.5 5.8 6
6 PH 7.2 7.3 7.2
7 TDS 590 510 480

The amount of calcium carbonate deposited on the membrane surface was determined
by mass balance considerations as a difference between the amount of calcium in the feed
water tank in the beginning of experiment and the amount of calcium carbonate at the
moment of the experiment. Deposition rates of calcium carbonate were defined as the
derivative of the function of the amount of deposited calcium over time.
Feed water (ground water in the Moscow region) was added to feed water tank 1
and was then pumped by pump 2 to membrane module 3. Membrane modules model
1812 70 NE with nanofiltration membranes and 1812 BLN with low pressure reverse osmosis
membranes were used. The Procon rotary pump was supplied by Procon Products, Smyrna,
TN, USA and produced 180–200 L per hour at a pressure of 16 bar. The experiments were
carried out using serial membrane elements of the 1812 standard model produced by Toray
Advanced Materials Korea Inc. (the manufacturer of CSM Membrane Technologies, Korea,
Seoul company CSM) with reverse osmosis membranes of the BLN model (selectivity for
salts of 95–96%) and nanofiltration elements of the model with membranes of the 70 NE
type with a selectivity of 70%. The area of the membranes in the apparatus model 1812 was
0.5 square meters.
Concentrate samples were taken from tank 1. Calcium, chloride, sulphate and bicar-
bonate ions concentrations as well as pH and TDS values were determined in the samples.
The test procedure and calculation techniques to evaluate scaling rates in the presence of an-
tiscalants and without their addition has been discussed in a number of publications [1,12].
Figure 2a shows dependencies of calcium concentrations in tank 1 on the coefficient K
value (that is determined as a ratio of the initial volume in tank 1 to the concentrate volume
at the moment of the experiment). The amount of calcium carbonate deposited in the
membrane channel was calculated as the difference between calcium amount in tank 1 at
the beginning and at the end of each test run (Figure 2b). Figure 2c shows dependencies
of calcium carbonate deposited in the membrane module as a function of time. Scaling
rates were determined as tangent values of calcium carbonate amount M (in milliequiva-
lents) versus time T function and were expressed in milliequivalents per hour (Figure 2d).
 antiscalants and without their addition has been discussed in a number of publications
[1,12]. Figure 2a shows dependencies of calcium concentrations in tank 1 on the coefficient
K value (that is determined as a ratio of the initial volume in tank 1 to the concentrate
volume at the moment of the experiment). The amount of calcium carbonate deposited in
the membrane channel was calculated as the difference between calcium amount in tank
1 at the beginning and at the end of each test run (Figure 2b). Figure 2c shows dependen-
Membranes 2023, 13, 85 4 of 15
cies of calcium carbonate deposited in the membrane module as a function of time. Scaling
rates were determined as tangent values of calcium carbonate amount M (in milliequiva-
lents) versus time T function and were expressed in milliequivalents per hour (Figure 2d).
Scaling
Scaling rates
rates weredetermined
were determined forforantiscalant
antiscalant doses
doses2, 52,
and 10 ppm
5 and for both
10 ppm forantiscalants
both antiscalants
(Figure 2d). During experimental test runs, antiscalant adsorption ratesalso
(Figure 2d). During experimental test runs, antiscalant adsorption rates were wereevalu-
also evalu-
ated. Figure 3 shows the results of determination of antiscalant adsorption rates to crystal
ated. Figure 3 shows the results of determination of antiscalant adsorption rates to crystal
surface. Adsorption rates were determined in accordance with the previously developed
surface. Adsorption
and described rates were
procedure [1,2]. determined
The rejection in accordanceof
characteristics with the previously
membranes developed
also signifi-
andcantly
described procedure [1,2]. The rejection characteristics of membranes
influence scaling propensities of membrane modules [13–16]. Figure 3 demon- also significantly
influence
strates scaling
results ofpropensities
evaluation ofof membrane
calcium modules
carbonate scaling[13–16]. Figure 3 demonstrates
rates in membrane 1812 elements results
of evaluation
tailored withofnanofiltration
calcium carbonate scaling rates
70NE membranes. As in membrane
is shown 18123b,
on Figure elements tailored
scaling rates in with
nanofiltration
nanofiltration membrane
70NE elementsAs
membranes. areis at least 4–5
shown on times
Figurelower than in rates
3b, scaling reverse
inosmosis
nanofiltration
elements.elements are at least 4–5 times lower than in reverse osmosis elements.
membrane

Membranes 2023, 13, x FOR PEER REVIEW 5 of 17

(a) (b)

(c) (d)
Figure 2. Evaluation of calcium carbonate scaling rates in membrane elements with reverse osmosis
Figure 2. Evaluation of calcium carbonate scaling rates in membrane elements with reverse osmosis low
low pressure membranes: dependences of concentrations of calcium in concentrate volume on K
pressure membranes:
values dependences
(a); dependencies of concentrations
of accumulated of calcium
calcium carbonate in concentrate
amount volume on of
on K (b); dependencies K cal-
values (a);
dependencies of accumulated
cium carbonate calcium
amount on time carbonateT amount
of experiment on K (b);
(c);dependencies ofdependencies of calcium
calcium carbonate carbonate
scaling rates
on on
amount K (d);
time1—Aminat-K, dose
of experiment 10 ppm; 2—Aminat-K,
T (c);dependencies dose 5carbonate
of calcium ppm; 3—Jurby-Soft, dose
scaling rates on10K ppm; 4—
(d); 1—Aminat-
Jurby-Soft, dose 5 ppm; 5—without antiscalant.
K, dose 10 ppm; 2—Aminat-K, dose 5 ppm; 3—Jurby-Soft, dose 10 ppm; 4—Jurby-Soft, dose 5 ppm;
5—without antiscalant.
 Figure 2. Evaluation of calcium carbonate scaling rates in membrane elements with reverse osmosis
low pressure membranes: dependences of concentrations of calcium in concentrate volume on K
values (a); dependencies of accumulated calcium carbonate amount on K (b); dependencies of cal-
cium carbonate amount on time of experiment T (c);dependencies of calcium carbonate scaling rates
Membranes 2023, 13, 85 on K (d); 1—Aminat-K, dose 10 ppm; 2—Aminat-K, dose 5 ppm; 3—Jurby-Soft, dose 10 ppm; 4— 5 of 15
Jurby-Soft, dose 5 ppm; 5—without antiscalant.

Membranes 2023, 13, x FOR PEER REVIEW 6 of 17

(a)

(b)
Figure 3. Results
Figure of evaluation
3. Results of calcium
of evaluation carbonate
of calcium scaling rates
carbonate in membrane
scaling elements elements
rates in membrane with nan-with
ofiltration membranes: dependencies of concentration of calcium in concentrate volume on
nanofiltration membranes: dependencies of concentration of calcium in concentrate volume K valueson K
(a); dependencies of calcium carbonate scaling rates on K values (b); 1—Aminat-K, 5 ppm; 2—Jurby-
values (a); dependencies of calcium carbonate scaling rates on K values (b); 1—Aminat-K, 5 ppm;
Soft, 5 ppm; 3—without antiscalant addition; 4—Aminat-K, 2 ppm; 5—product flow after addition
2—Jurby-Soft, 5 ppm; 3—without antiscalant addition; 4—Aminat-K, 2 ppm; 5—product flow after
of 5 ppm of Aminat-K; 6—product flow after addition of 2 ppm of Aminat-K; 7—product flow with-
addition of 5 ppm of Aminat-K; 6—product flow after addition of 2 ppm of Aminat-K; 7—product
out antiscalant addition.
flow without antiscalant addition.
During the conducted test runs, concentration values of antiscalants (concentrations
During the conducted test runs, concentration values of antiscalants (concentrations of
of phosphate ions) were determined (Figure 4a). Dependencies of antiscalant adsorption
phosphate ions) were determined (Figure 4a). Dependencies of antiscalant adsorption rate
rate values
values on
on K
K values
values are
are shown
shown on
on Figure
Figure 2b.
2b. The
The higher
higher the
the antiscalant
antiscalant dose,
dose, the
thehigher
higher the
the adsorption
adsorption rate.
rate.AArelationship
relationship between
between adsorption
adsorption rates
rates and and antiscalant
antiscalant dosesdoses is in
is shown
shown in Figure 5. These relationships are built for the K value of 5. These relationships
demonstrate the comparison of properties of different antiscalants.
 Membranes 2023, 13, 85 6 of 15

Membranes 2023, 13, x FOR PEER REVIEW 7 of 17

Figure 5. These relationships are built for the K value of 5. These relationships demonstrate
the comparison of properties of different antiscalants.

(a)

(b)
Figure 4. Dependencies
Figure 4. of Aminat-K and
Dependencies ofJurby-Soft
Aminat-Kdoses
and in RO concentrate
Jurby-Soft on RO
doses in K values (a) and on K values (a) and
concentrate
dependencies of adsorption rate values on K (b): 1—Aminat -K, dose 2 ppm; 2—Aminat-K, dose 2
dependencies of adsorption rate values on K (b): 1—Aminat -K, dose 2 ppm; 2—Aminat-K, dose 2 ppm;
3—Aminat-K, dose 10 ppm; 4—Jurby-Soft, 2 ppm; 5—Jurby-Soft, 5 ppm; 6—Jurby-Soft, 10 ppm.
Membranes 2023, 13, x FOR PEER REVIEW 8 of 17

Membranes 2023, 13, 85

ppm; 3—Aminat-K, dose 10 ppm; 4—Jurby-Soft, 2 ppm; 5—Jurby-Soft, 5 ppm; 6—Jurby-Soft, 10 7 of 15
ppm.

Figure 5. Dependencies of antiscalant adsorption rates on the antiscalant dose values: (1)—Aminat-K;
Figure(2)—Jurby-Soft.
5. Dependencies of antiscalant adsorption rates on the antiscalant dose values: (1)—Aminat-
K; (2)—Jurby-Soft.
3. Experimental Results
3. Experimental
3.1. Effect ofResults
the Antiscalant Dose on Scaling Rates
3.1. Effect ofFigure
the Antiscalant
2 shows theDosemain
on Scaling
steps inRates
determining calcium carbonate scaling rates in the
presence of antiscalants and without antiscalant
Figure 2 shows the main steps in determining calcium addition. The amount
carbonate of calcium
scaling rates incarbonate
the
wasof
presence determined
antiscalantsby and
a mass balance
without as a difference
antiscalant between
addition. Thethe amount
amount ofofcalcium
calciumcar-
in tank 1
bonate was determined by a mass balance as a difference between the amount of calciumexper-
at the beginning of the experiment and calcium amount in tank 1 at the moment of the
in tankiment [6].beginning
1 at the Figure 2a demonstrates
of the experimentdependencies
and calcium of deposited
amount in calcium
tank 1 atcarbonate
the moment amounts
on coefficient K values. Calcium carbonate deposition rates were
of the experiment [6]. Figure 2a demonstrates dependencies of deposited calcium car- determined in conformity
bonatewith the method
amounts described
on coefficient K in [6] as Calcium
values. tangents carbonate
of the function of calcium
deposition ratescarbonate
were deter- amount
versus time (Figure 2b). Results of scaling rate determination are
mined in conformity with the method described in [6] as tangents of the function of cal- demonstrated in Figure 2c
as dependencies of calcium carbonate scaling rate values (expressed as milliequivalents
cium carbonate amount versus time (Figure 2b). Results of scaling rate determination are
of calcium carbonate per hour) on coefficient K value. Figure 2c demonstrates coefficient
demonstrated in Figure 2c as dependencies of calcium carbonate scaling rate values (ex-
K values that correspond to beginning of scaling in the RO module under different condi-
pressed as milliequivalents of calcium carbonate per hour) on coefficient K value. Figure
tions in the presence of antiscalants and without antiscalant addition. Different K values
2c demonstrates coefficient K values that correspond to beginning of scaling in the RO
correspond to different scale formation (nucleation) conditions in the “dead” area [2].
module under different conditions in the presence of antiscalants and without antiscalant
addition.
3.2. Different K valuesType
Effect of Membrane correspond
on ScalingtoRates
different scale formation (nucleation) condi-
tions in the “dead” area [2].
Aminat-K demonstrates lower scaling rates than Jurbi-Soft at different doses. The
membrane type also influences scaling rate. As can be seen in Figure 3, scaling rates in the
3.2. Effect of Membrane Type on Scaling Rates
nanofiltration membrane element are substantially lower than in the membrane element
Aminat-K
with reverse demonstrates lower scaling
osmosis membranes underratesallthan
the Jurbi-Soft at differentconditions:
same experimental doses. Thewater
membrane type also influences scaling rate. As can be seen in Figure 3, scaling
composition, pressure, recovery, antiscalant dose. Reduced scaling rates in nanofiltration rates in the
nanofiltration membrane element are substantially lower than in the
membrane modules are attributed to low rejection values and lower supersaturation membrane element
with reverse
conditionsosmosis
in themembranes
“dead” areasunderin all the same channels
membrane experimental conditions:
which creates thewater com- for
conditions
position, pressure,
the start recovery, antiscalant
of crystallization [1,12]. Thedose.
use ofReduced scalingmembranes
low rejection rates in nanofiltration
in drinking water
membrane modules
production are attributed
projects to low rejection
can be considered alongvalues and lower supersaturation
with antiscalant con-
addition as a measure to
ditionscontrol
in thescaling
“dead”[2].areas in membrane channels which creates the conditions for the
Membranes 2023, 13, 85 8 of 15

3.3. Evaluation of Antiscalant Adsorption Rates and Their Influence on Scaling Rates
Figures 4 and 5 demonstrate results of Aminat-K and Jurby-Soft adsorption rate
measurements during calcium carbonate deposition on membrane surface. Aminat-K
exhibits higher adsorption abilities than Jurby-Soft at the same doses. Stronger Aminat-K
provides higher adsorption rate and higher calcium carbonate scaling rate reduction at
10 ppm dose when coefficient K equals 1.5, but Jurby-Soft cannot provide such an efficiency
value and such an adsorption rate with a 10 ppm dose. Figure 5 shows that Aminat-K is a
strong antiscalant and even at a small dose of 2 ppm provides higher adsorption rates than
Jurby-Soft. This also explains why Aminat-K shows the same results in the concentration
range from 2 to 7 ppm while Jurby-Soft gives different results under the same conditions.

3.4. Selecting Antiscalant and Membrane Type to Reduce Operational Costs

Antiscalant suppliers are usually limited by general recommendations to add 1 to
10 milligrams of antiscalant per one liter of RO feed water. More detailed recommendations
can be provided in accordance with the calculations that suppliers have developed for
their product for a variety of application conditions. Thus, the dosing of antiscalant
changes depending on supersaturation ratio reached in the membrane module depending
on feed water composition and recovery values [1]. Calculations often give inconsistent and
controversial results as they depend on the experimental conditions under which the crystal
formation and scale deposition were conducted [1,2]. In our experiments we evaluate the
scaling rates using the ground water from the source, industrial membrane module and
pressure and recovery conditions relevant to the industrial operation conditions of the RO
plant. This enables us to provide a rapid quantitative evaluation of scale deposition rate
in the presence of different amounts of antiscalant and to find an answer: which makes it
possible to reduce the dose without compromising the effectiveness of scaling control.
Application of antiscalants provides reduction of scaling and higher values in the
period between the cleanings. Figures demonstrate results of scaling rate evaluations as a
function of coefficient of initial volume reduction K for the cases of different antiscalants
and different dose applications. Figure 6a shows the dependence of accumulated calcium
carbonate amount on time of operation that was obtained previously as a result of inves-
tigations [2,16] using 1812 elements tailored with low pressure RO and NF membranes.
These relationships are used for prognosis of product flow and rejection. Figure 6b shows
the recommended time for membrane cleaning for the presented case study; time periods
T1, T2 and T3 correspond to the exact amount of calcium carbonate (600 milliequivalents)
deposited on 1 square meter of membrane.
Table 1 shows the main ground water compositions met in water intake in the Moscow
region. Hardness can vary between 5 and 7 milliequivalents and alkalinity between 4 and
6 milliequivalents per liter. To predict antiscaling behaviour of Aminat-K antiscalant in
a variety of conditions, authors have developed software to help RO operators to select
recoveries and cleaning schedules. For three different feed water compositions presented
in Table 1, results of calcium carbonate scaling rate evaluation are presented (Figure 7a) as
a function of coefficient K value and recommended time of operation between cleanings
is calculated (Figure 7b). For cases when a different new antiscalant (Jurby-Soft) is used,
Figure 7 shows results of scaling rates and time period between cleanings as a comparison.
In cases when another antiscalant is used which demonstrates lower efficiency, the
time period between cleanings decreases. Very often. applied cleanings are economically
unreasonable. If we extend the period between cleanings, the amount of accumulated
calcium carbonate grows and the efficiency of membrane cleaning decreases [2,16]. Thus, a
certain amount of scale is left after routine cleanings and constantly grows.
Membranes 2023, 13, x FOR PEER REVIEW 10 of 17
Membranes 2023, 13, 85 9 of 15

(a)

(b)
Figure 6. Prediction
Figure of membrane
6. Prediction of membrane product
productflow
flowdecrease
decreasewith
withcalcium
calciumcarbonate
carbonate accumulation
accumulation over
time andtime
over recommended time between
and recommended membrane
time between membranecleanings: (a)—dependencies
cleanings: (a)—dependencies ofof accumulated
accumulated cal-
cium carbonate
calcium amount
carbonate on on
amount time; (b)(b)dependencies
time; of product
dependencies of productflow
flowonon time;
time; 1—Jurby-Soft,
1—Jurby-Soft, 5 ppm;
5 ppm;
2—Jurby-Soft, 10 ppm; 3—Aminat-K, 5 ppm; 4—Aminat-K,
2—Jurby-Soft, 10 ppm; 3—Aminat-K, 5 ppm; 4—Aminat-K, 10 ppm. 10 ppm.

Table 1 shows the main ground water compositions met in water intake in the Mos-
cow region. Hardness can vary between 5 and 7 milliequivalents and alkalinity between
4 and 6 milliequivalents per liter. To predict antiscaling behaviour of Aminat-K antiscalant
in a variety of conditions, authors have developed software to help RO operators to select
recoveries and cleaning schedules. For three different feed water compositions presented
in Table 1, results of calcium carbonate scaling rate evaluation are presented (Figure 7a)
as a function of coefficient K value and recommended time of operation between cleanings
is calculated (Figure 7b). For cases when a different new antiscalant (Jurby-Soft) is used,
Figure 7 shows results of scaling rates and time period between cleanings as a comparison.
Membranes
Membranes2023,
2023,13,
13,x 85
FOR PEER REVIEW 11 10
of of
1715

(a)

(b)
Figure
Figure7. 7.
Prediction of membrane
Prediction of membraneproduct flowflow
product decrease with calcium
decrease carbonate
with calcium accumulation
carbonate over
accumulation
time
overand recommended
time time between
and recommended membrane
time between cleanings:
membrane (a)—dependencies
cleanings: of accumulated
(a)—dependencies cal-
of accumulated
cium carbonate
calcium amount
carbonate on time;
amount (b) (b)
on time; dependencies of product
dependencies flow
of product flowonon
time; 1—water
time; 1—water composition
composition
1, Aminat K 5; 2—water composition 1, Jurby Soft 5 ppm; 3—water composition 2, Aminat K 5; 4—
1, Aminat K 5; 2—water composition 1, Jurby Soft 5 ppm; 3—water composition 2, Aminat K 5;
water composition 2, Jurby Soft 5 ; 5—water composition 3, Aminat K 5; 6—water composition 3,
4—water composition 2, Jurby Soft 5; 5—water composition 3, Aminat K 5; 6—water composition 3,
Jurby Soft 5 ppm; 7—water composition 1, without antiscalant addition (Table 1).
Jurby Soft 5 ppm; 7—water composition 1, without antiscalant addition (Table 1).
Membranes 2023, 13, 85 11 of 15

4. Industrial Application of the Results

Antiscaling efficiency of the dosed antiscalant provides reduction in scaling rates and
longer operational period between cleanings. Figure 6 shows the results of scaling rates
determination as a dependence on K value The plots shown in Figure 6a are used for
Membranes 2023, 13, x FOR PEER REVIEW
prognostic techniques to recommend operational time for predicted product flux and13rejec- of 17
tion decrease, based on results of experimental evaluation of scaling rates [1,2]. Figure 6b
recommends an operational time value of 300 h for the case of ground water composition
11 Number of cleanings and per
K value
yearof 4. For cases when different8 antiscalants are used with higher 4 or lower scaling
rates and retardation efficiencies, operational time required to reach cleaning should be
Reagent consumption;
either shortened or expanded. In our case, replacing of Jurby-Soft antiscalant by Aminat-K
12 Caustic cost, USD/liter, 6000/100 3000/100
provides nearly 50% reduction of scaling rates and therefore ensures safe operation of the
Acid cost, USD/liter 4800/120 2400/120
reverse osmosis facility within 600 h instead of 300. As discussed earlier (Figure 2a), Aminat-
Annual cleaning costs;
K provides efficient scaling rate reduction even using reduced dose values (1.5–2 ppm).
13 caustic, USD
Thus, total annual operational cost8571 savings due to application of Aminat-K 4286include reduced
acid, USDannual consumption of cleaning reagents 8229 and reduced consumption of4114 antiscalant. Table 2
14 Membranes; model/number/cost, USD Lewabrane RO B 370 HF/108 500
describes the calculation steps for evaluation of annual operational costs for the reverse
Annual costs for membrane osmosisreplace-
plant operated in the Moscow region with capacity of 130 cubic meter per hour.
15 9000 9000
ment, USDThe plant consists of two parallel lines, each including nine pressure vessels with six mem-
16 Annual operational brane
costs,8040
USD elements (Figure 7). Figures
49,400 8 and 9 describe the membrane vessels’ array and
31,400
17 the
Water cost price, USD/m graph shows the increase of K while
0.097 water is moved from the first element
0.05 to the last,
or “tail”
Membrane elements to reduce concen- element, as well as scaling rate values. This helps us to evaluate scaling rates in
18 each membrane element and total scaling 8040
rate and 70 NE/24/750
total amount of calcium carbonate scale
trate flow: model/number/cost, USD
accumulated
Additional annual costs for membrane in the array (Figure 6). The amount of required cleaning reagent (usually
19 citric acid and EDTA) and cleaning-schedules are taken according to results 3000 of cleanings.
replacement, USD
Figure 10 shows results of routine cleanings applied every 300 h of continuous operation
Concentrate discharge flow,
20 (on average once a month) with citric - acid. Assuming volume of cleaning 0.33 tank as 3 cubic
m /hour/tax, USD/m
meters, it can be calculated that an amount of 1500 equivalents can be removed using three
Annual payment for concentrate dis- Citric acid concentration was 3% (Figure 11).
cleaning series.
21 70,000 3500
charge, USD

Figure 8. A 120 cubic meter per hour reverse osmosis membrane plant for production of drinking
Figure 8. A 120 cubic meter per hour reverse osmosis membrane plant for production of drinking
quality
qualitywater
waterfrom
fromthe
theground
groundwater
waterwell
wellintake
intakeininthe
theMoscow
Moscowregion.
region.
Membranes 2023, 13, 85 12 of 15

Table 2. Evaluation of operational costs: economical comparison of antiscalant efficiencies and

recovery increase due to application of nanofiltration.

Present Technique Recommended Technique

Characteristics
(Antiscalant Dose 6 ppm) (Antiscalant Dose 4/2 ppm)
1 Feed flow, (m3 /H) 130/910,000 130/910,000
Product flow, 100 100
2 Designed, (m3 /H)/ 700,000/ 700,000/
Real, (m3 /year) 520,000 630,000
3 Concentrate flow, m3 /H 30 30/1.5 (After recovery increase using NF)
4 Pressure, Bar 14–16.5 14–15
5 Annual power consumption, KW 112,000 105,000
6 Power costs, USD/year 8000 7500
7 Antiscalant type, and dose, ppm Jurbysoft, 6 ppm Aminat-K, 4/2
8 Annual antiscalant consumption, kg 5460 3640/1820
9 Antiscalant costs, USD per ton 2857 3571
10 Annual antiscalant cost, USD 15,600 13,000/6500
11 Number of cleanings per year 8 4
Reagent consumption;
12 Caustic cost, USD/liter, 6000/100 3000/100
Acid cost, USD/liter 4800/120 2400/120
Annual cleaning costs;
13 caustic, USD 8571 4286
acid, USD 8229 4114
14 Membranes; model/number/cost, USD Lewabrane RO B 370 HF/108 500
15 Annual costs for membrane replacement, USD 9000 9000
16 Annual operational costs, USD 49,400 31,400
17 Water cost price, USD/ m3 0.097 0.05
Membrane elements to reduce concentrate flow:
18 8040 70 NE/24/750
model/number/cost, USD
19 Additional annual costs for membrane replacement, USD - 3000
20
Membranes Concentrate discharge
2023, 13, x FOR flow, m3 /hour/tax, USD/m3
PEER REVIEW - 0.33 14 of 17
21 Annual payment for concentrate discharge, USD 70,000 3500

(a)

(b)
Figure
Figure9. 9.
Evaluation of K
Evaluation ofvalues in membrane
K values elements
in membrane of theof
elements RO plant
the RO array:
plant (a) dependencies
array: of
(a) dependencies of K
K values on the membrane stage number; (b) pressure vessels array in membrane plant.
values on the membrane stage number; (b) pressure vessels array in membrane plant.
 (b)

Membranes 2023, 13, 85 Figure 9. Evaluation of K values in membrane elements of the RO plant array: (a) dependencies
13 ofof15
K values on the membrane stage number; (b) pressure vessels array in membrane plant.

(a)

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s)
and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or
property resulting from any ideas, methods, instructions or products referred to in the content.

land-2053789

xml:

1. keep order of aff address;

2. changed refs 16,19,43;

3. keep font format of n; (b)

4. keep format of Highlights, with special
Figure approval;
10. Evaluation of calcium carbonate scaling rates in membrane elements of the RO plant
5. table/figure without extra explanation, confirmed by PE.of scaling rates on the membrane stage number; (b) pressure vessels array in
array: (a) dependencies
membrane plant; 1—Jurby-Soft, 5 ppm; 2—Aminat-K, 2 ppm.

Figure 11. Dependencies of calcium ionic concentration in citric acid cleaning solution on time:
1—first cleaning; 2—second cleaning; 3—third cleaning.
Membranes 2023, 13, 85 14 of 15

Another expense item is concentrate disposal. Concentrate produced by a reverse

osmosis plant used for drinking water supply is forwarded to a sanitation sewer. For cases
when ground water in Moscow region is treated by RO, the K value is usually 3–4, which
means that concentrate flow value is 1/3 to 1/4 of the feed water flow. Thus, the water user
water tax includes not only the costs for water treatment but costs for additional wastewater
discharge. To reduce concentrate flow, an additional stage of nanofiltration membranes is
used [1,16]. Concentrate enters nanofiltration membrane modules where it is separated
into the product stream and concentrate stream that equals 1/10–1/20 of the feed water
flow. Nanofiltration membrane produce water is added to the feed ground water and this
increases recovery and the amount of water taken from the well and pretreated. Figure 10
shows the flow diagram of membrane modules array and the graph shows scaling rate
increase due to K value growth. When purchasing large amounts of antiscalant to avoid
fakes, it seems reasonable to apply analysis to the samples of the product to be sure that
the product has required MIDF concentration.

5. Conclusions
1. Antiscaling efficiency of the inhibitor depends on its ability to adsorb on the surface
of growing crystals. The higher adsorption rate, the better the antiscaling behavior of
the inhibitor.
2. To efficiently control scaling and to reduce operational costs of the RO unit it is
recommended to tailor the membrane plant with nanofiltration membranes. The joint
use of low rejection membranes and efficient antiscalant provides substantial decrease
in scaling rates in membrane modules and reduces operational costs.
3. High adsorption abilities of phosphonic-based antiscalant enables us to reduce the
antiscalant dose in the feed water without compromising effectiveness of scale control
and thus reduce reagent consumption and operational costs.
4. High efficiency of phosphonic antiscalants is attributed to the content of aminotris (methy-
lene=phosphonic acid) in the product. Application of Nuclear Magnetic Resonance method
helps to identify the presence of ATMP and to avoid buying low-quality products.

Author Contributions: Conceptualization, D.S. and A.P.; methodology, A.P.; software, H.Z.A.; vali-
dation, D.S. and A.P.; formal analysis, D.S.; investigation, D.S.; resources, D.S.; data curation, H.Z.A.;
writing—original draft preparation, H.Z.A.; writing—review and editing, A.P.; visualization, H.Z.A.;
supervision, D.S.; project administration, D.S.; funding acquisition, D.S. All authors have read and
agreed to the published version of the manuscript.
Funding: This work was financially supported by the Ministry of Science and Higher Education of
Russian Federation (grant # 075-15-2021-686). Tests were carried out using research equipment of the
Head Regional Shared Research Facilities of the Moscow State University of Civil Engineering.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available on request from the
corresponding author.
Acknowledgments: The scientific consult and assistance of Galina Rudakova (Materials) is gratefully
acknowledged.
Conflicts of Interest: The author declares no conflict of interest.

References
1. Pervov, A.; Andrianov, A. Deposition of Calcium and Magnesium from RO concentrate by means of seed crystallization and
production of softened water for technical purpose. Desalination Water Treat. 2018, 110, 10–18. [CrossRef]
2. Pervov, A.G.; Nguen, X.Q. Application of reverse osmosis and nanofiltration techniques at municipal drinking water facilities.
E3S Web Conf. 2019, 97, 06004. [CrossRef]
3. Suratt, W.B.; Adrews, D.R.; Pujals, V.J.; Richards, S.A. Design considerations for major membrane treatment facility for groundwa-
ter. Desalination 2000, 131, 37–46. [CrossRef]
Membranes 2023, 13, 85 15 of 15

4. Salman, M.A.; Al-Nuwaibit, G.; Safar, M.; Al-Mesri, A. Performance of physical treatment method and different commercial
antiscalants to control scaling deposition in desalination plant. Desalination 2015, 356, 294–313. [CrossRef]
5. Turek, M.; Mitko, K.; Dydo, P.; Laskovska, E.; Jakobic-Kolon, A. Prospects for high water recovery membrane desalination.
Desalination 2017, 401, 180–189. [CrossRef]
6. Goh, P.S.; Lau, W.J.; Othman MH, D.; Ismail, A.F. Membrane fouling in desalination and its mitigation strategies. Desalination
2018, 425, 130–155. [CrossRef]
7. Jiang, S.; Li, Y.; Ladewig, B.P. A review of reverse osmosis membrane fouling and control strategies. Sci. Total Environ. 2017, 95,
567–583. [CrossRef] [PubMed]
8. Li, X.; Hasson, D.; Shemer, H. Flow conditions affecting the induction period of CaSO4 scaling on RO membranes. Desalination
2018, 431, 119–125. [CrossRef]
9. Greenlee, L.F.; Lawler, D.F.; Tesla, F.; Freeman, B.D. The effect of antiscalant addition on calcium carbonate precipitation for a
simplified synthetic brackish water reverse osmosis concentrate. Water Res. 2010, 44, 2957–2969. [CrossRef] [PubMed]
10. Shahid, M.K.; Choi, Y.-G. The comparative study for scale inhibition on surface of RO membranes in wastewater different
antiscalants. J. Membr. Sci. 2018, 546, 61–69. [CrossRef]
11. Chaussemier, M.; Pourmohtasham, E.; Gelus, D.; Pecoul, N.; Perrot, H.; Ledion, J.; Cheap-Charpentier, H.; Homer, O. State of art
of natural inhibitors of calcium carbonate scaling: A review article. Desalination 2015, 356, 47–55. [CrossRef]
12. Pervov, A.; Andrianov, A.; Rudakova, G.; Popov, K. A comparative study of some novel “green” and traditional antiscalants
efficiency for the reverse osmotic Black Sea water desalination. Desalination Water Treat. 2017, 73, 11–20. [CrossRef]
13. Zimmer, K.; Hater, W.; Icart, K.; Jaworski, J.; Kruse, N.; Braun, G. The performance of polycarboxylates as inhibitors for CaCO3
scaling in reverse osmosis plants. Desalination Water Treat. 2016, 57, 48–49. [CrossRef]
14. Al-Qadami, E.; Ahsan, A.; Mustafa, Z.; Abdurrasheed, A.; Yusof, K.; Shah, S. Nanofiltration membrane technology and its
applications in surface water treatment: A review. J. Desalination Water Purif. 2020, 18, 3–9. Available online: http://ababilpub.
com/download/jdwp18-2 (accessed on 27 December 2022).
15. Guo, H.; Li, X.; Yang, W.; Yao, Z.; Mei, Y.; Peng, L.E.; Yang, Z.; Shao, S.; Tang, C.Y. Nanofiltration for drinking water treatment: A
review. Front. Chem. Sci. Eng. 2021, 16, 681–698. [CrossRef] [PubMed]
16. Mohammad, A.W.; Hilal, N.; Darwish, N.; Al-Zoubi, H. Prediction of permeate fluxes and rejections of highly concentrated salts
in nanofiltration membranes. J. Membr. Sci. 2007, 289, 40–50. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.